Professor R.M. Hare is well-known both for his development, over the years, of a distinctive moral theory, and for his interest in questions of practical morality, among which moral education has been prominent [1]. His recent book, Moral Thinking [2], presents the latest version of his theory and draws out its implications for our practical reasoning; it will be particularly interesting to educators and philosophers of education because of the importance Hare now lays on there being two levels of moral thinking. Moral education, according to Hare, can only proceed in a clear and unconfused way when we have learnt to distinguish these levels, for they correspond to different levels of moral development. Hare's moral theory is, he now claims, a form of utilitarianism; and both this and his separation of levels of moral thinking are worth studying in some detail, since Hare is in tune with a strong tradition of recent moral philosophy in claiming that morality requires a theory, that that theory must be utilitarian, and that our pursuit of utility must be on more than one level. Hare has always insisted that, if I am to make a moral judgement (as opposed to an aesthetic, or prudential one) I must be prepared to universalise it. This is a matter not of the content of the judgement but of the type of reason offered for it; the reasons offered for a moral judgement have to be independent of the way I happen to be, and to apply with equal force to anyone in a relevantly similar situation. The basic idea is the Kantian one that morality applies to all agents regardless of personal differences; a moral reason for me is a moral reason for anyone in a relevantly similar situation. In Moral Thinking Hare expresses the universalisability claim, as he has done before, as a claim about the agent's consistency: if I make a moral judgement in one situation, I cannot refuse to make the same judgement in a relevantly similar situation-or at least I can only at the cost of inconsistency. (We might comment that some people manage to live with a fair degree of inconsistency in their beliefs, but Hare does not treat this as a real option.) In previous work, however, up to and including the well-known Freedom and Reason, Hare joined to this thesis another, that of the essential prescriptivity of moral judgements, which left it open to the agent to adopt moral principles of unrestricted content, provided they respected the consistency imposed by the universalisability requirement. Thus, notoriously, the problem of the fanatic: a clear-headed Nazi could on this view count as holding, as a moral principle, the principle that Jews ought to be exterminated, provided he or she were not making exceptions in their own interest and were therefore prepared to stick to the principle even in the hypothetical situation of turning out to be a Jew. Hare's insistence that we are free to commit ourselves to any principle and hold it as 199